Background
Scrub typhus is a rickettsial disease caused by
Orientia tsutsugamushi [
1], which is transmitted to humans through infected chigger mites. Scrub typhus is widely distributed in Southeast Asia and the Pacific Rim including China [
1,
2]. When the rickettsia is transmitted through the bite of an infected mite to human, it begins to proliferate at the bite site and a characteristic skin lesion, known as an eschar, is formed. The pathogen then spreads systemically via the hematogenous and lymphogenous routes. Infected people develop various systemic symptoms and reactions including fever, cutaneous rash, lympadenopathy, elevations of C-reacting protein (CRP) and liver enzymes [
2‐
4].
Prior to 1986, scrub typhus was only found endemic in southern China (south of the Yangtse River, or to the south of 31° north latitude), including 11 provinces (e.g., Guangdong, Hainan, Guangxi, Fujian, Zhejiang, Yunnan, Hunan province). Because human infections typically occur between March and November with a peak occurrence between June and August in the summer [
5‐
7], so the scrub typhus is also called "summer type" scrub typhus, which is transmitted by the
Leptotrombidium deliense mite [
5‐
8]. The reservoir hosts are rodents mainly including
Rattus losea,
R. flavipectus, and
Apodemus agrarius [
5‐
7]. Major serotypes of summer type scrub typhus in many areas of southern China were Karp, Gilliam, and Kato types [
5‐
7,
9‐
11]. In China, genotypes of scrub typhus have not been systematically studied until recently [
6,
7,
9]. However, the genotyping results obtained in Guangdong, Fujian, Hainan province of southern China revealed that Karp types were the key genotypes of summer type scrub typhus in these areas [
12‐
14]. The summer type scrub typhus is caused by a relatively more virulent strain of
O. tsutsugamushi [
5,
7]. Human cases caused by the summer type scrub typhus have common clinical features including fever, cutaneous rash, eschar and local lympadenopathy, and the associated complications were typically severe [
5‐
7].
During the autumn-winter period of 1986, some residents in Mengyin county, south of Shandong province [
15], and Dongtai, northern Jiangsu province [
16] (both located in north of the Yangtse River, or to the north of 31° north latitude) developed an unknown fever, which was later identified to be caused by scrub typhus. As cases associated with this type of scrub typhus occurred from September to December with an occurrence peak in October, it was called "autumn-winter type". This type of scrub typhus was subsequently reported in many regions of northern China including Tianjing, Shanxi province, Hebei province, and Henan province [
5‐
7,
17,
18]. The autumn-winter type scrub typhus is caused by a less virulent strain of
O. tsutsugamushi and transmitted by
L. scutellare mite [
5‐
7,
19]. The reservoir hosts are
A. agrarius,
Cricetulus triton, and
R. norvegicus [
5‐
7,
19]. Although Gilliam types were identified by IFA as the key serotypes of autumn-winter type scrub typhus in many areas of northern China [
5‐
7,
14,
19], the genotyping results acquired in Shandong, northern Jiangsu province in northern China showed that Kawasaki types were the key genotypes of this type of scrub typhus in the study areas (the genotypes in other areas were not studied or undetermined) [
14,
20‐
22]. Since 1986, people infected with autumn-winter type scrub typhus have been increasing in many areas of northern China [
5‐
7,
17,
18]. However, clinical characteristics of the newly recognized scrub typhus have not been studied. In this study, we summarize the clinical characteristics of human infections by autumn-winter type scrub typhus in Feixian county, south of Shandong province, northern China.
Discussion
In this study, we documented occurrence periods for the autumn-winter type scrub typhus cases. The seasonal variation of scrub typhus cases in northern China was similar to that reported in Korea, where human cases of scrub typhus were reported to increase in October and peak in November [
29]. This pattern was, however, different from those reported in Japan [
30,
31] and in southern China [
5‐
7]. In Japan, a bimodal pattern of occurrence of cases – one in spring and another in autumn-winter (the latter being similar to the autumn-winter type reported in current study) [
30,
31] – was reported. Yet in southern China, scrub typhus, caused by the summer type, was endemic with a single peak of infections occurring in summer [
5‐
7,
28]. Because humans are infected through bites of the larva of the chigger mites, seasonal variations in scrub typhus infections may be in part due to seasonal fluctuations of the larval chigger mites as well as their rodent hosts. In Japan, the bimodal pattern of occurrence of cases was reported to relate to population dynamics of the two different species of chigger mites. Cases occurred in the autumn-winter period in many areas of Japan were mainly due to
L. scutellare, while cases in the spring period in western Japan were caused primarily through
L. pallidum [
30,
31]. In southern China, seasonal variations in scrub typhus infections were commonly consistent with seasonal variations in population of
L. deliense [
5‐
8]. The authors conducted a survey on rodent dynamics between May 1995 and April 1996 in the same study area and found that the population abundance of
A. agrarius, the most abundant rodent species in the enzootic area, showed a bimodal pattern in a year – the first in July and the second in October.
O. tsutsugamushi was isolated from
A. agrarius, which was confirmed as the host for scrub typhus in the areas [
19,
32]. Seasonal variations in chigger mites on the field rodents were also observed in this one-year study, showing that
L. scutellare first appeared during the period of day 11–20 of September (the chigger index was 13.51) and its abundance then increased sharply in October (the chigger index increased to 22.68). The chigger population peaked during the period of day 11–20 of November (the chigger index peaked to 36.56) and dropped substantially in December (the chigger index in December was 0.47). In addition,
O. tsutsugamushi was also isolated from
L. scutellare, which was confirmed as the vector of scrub typhus in the study area [
19,
33]. The present study, together with our previous work [
19,
32,
33], suggests that the time when rodents cause maximum larval chigger infestation overlaps the period when the highest incidence of scrub typhus infections were reported in humans. Based on previous reports and this study, the main differences between the "summer type" and "autumn-winter type" scrub typhus in China were summarized in Table
6.
Table 6
Key differences between "summer type" and "autumn-winter type" scrub typhus in China
Seasonal distribution of human cases | Between March and November with a peak occurrence in the summer between June and August | Exclusively from September to December with a peak occurrence in October | |
Geographical distribution | Endemic in southern China (south of the Yangtse River), including Guangdong, Hainan, Guangxi, Fujian, Zhejiang, Yunnan, and Hunan province. | Endemic in northern China (north of the Yangtse River), including Shandong, northern Jiangsu, Tianjing, Shanxi, Hebei, and Henan province | |
Key reservoir hosts |
Rattus losea, R. flavipectus, and Apodemus agrarius
|
A. agrarius, Cricetulus triton, and R. norvegicus. | |
Key vector chigger mites |
Leptotrombidium deliense
|
L. scutellare
| |
Virulence of O. tsutsugamushi isolates
| More virulent, because: (1) Being tested with mice, the median lethal dose (LD50) values of O. tsutsugamushi isolates were 10-5 – 10-8; (2) O. tsutsugamushi could be successfully isolated from cases by normal experimental mice without any treatment | Less virulent, because: (1) Partial O. tsutsugamushi isolates do not lead to death of mice; (2) Being tested with mice, LD50 values of most O. tsutsugamushi isolates were 10-0.5 – 10-3; (3) In order to isolate O. tsutsugamushi successfully from cases, the experimental mice must be treated by cyclophosphamide to suppress immunity | |
Key serotypes | Karp type, Gilliam type, and Kato type | Gilliam type | |
Key genotypes* | Karp type | Kawasaki type | |
Clinical features and complications | Cases have the common clinical features of scrub typhus such as fever, cutaneous rash, eschar and local lympadenopathy. However, the associated complications of this type scrub typhus were severe | Cases also have the common clinical features of scrub typhus such as fever, cutaneous rash, eschar and local lympadenopathy. However, the associated complications of this type of scrub typhus were less severe than those of the summer type | |
Although certain clinical symptoms occurred more frequently in autumn-winter type cases than in summer type cases, regional lymphadenopathy, retro-orbital pain, electrocardiogram abnormity, hepatosplenomegaly, and flank tenderness were less common in the autumn-winter type cases. In contrast, some severe complications such as toxic myocarditis, alimentary tract hemorrhage, pleural fluid, or abdominal dropsy, commonly reported in the summer type cases in southern China [
28], were not found at all in present study (Table
4). In addition, the abnormalities of WBC and PLT counts in present study were also less obvious than those reported for the summer type. The reports in southern China suggested that the percentages of cases with the increased and decreased WBC counts were 32.9% and 27.0% respectively, which were higher than (χ
2 = 219.7,
P < 0.001) those reported in current study. Thrombocytopenia was reported in 51.1% cases in southern China (Table
5) [
28], also higher than (χ
2 = 222.7,
P < 0.001) that obtained in current study.
The reported percentages of eschar formation showed substantial variations across different studies [
34‐
37], which could be explained in part by disparity of physicians experiences. Eschars could be detected relatively frequently on white-skinned Japanese children, however, it is relatively difficult to detect eschars on dark-skinned Thai pediatric cases [
35‐
37]. Previous studies from some new endemic areas in northern China showed that the percentages of scrub typhus cases with eschars were 15% [
38], 34% [
39], 84% [
40] and 100% respectively [
18]. In our study, 88.5% of cases had eschars. This is relatively high compared to other reports. Part of the reason is that physicians in our study team have worked there for many years and were very familiar with eschars. In addition, in our study we carried out thorough body examination for each suspected scrub typhus case, which may have contributed to high detection rate of eschars.
Irons et al (1947) reported that 45% of confirmed eschars in US army personnel were detected on the feet and legs. Perineum, inguinal area, and axilla were also the preferentially eschar-manifested areas [
41]. Kim et al (2007) reported that among the 162 adult scrub typhus cases in southwestern area of Korea, most cases had eschars on the front of the body. Eschars were primarily detected in males within 30 cm below the umbilicus. Yet a different pattern was seen in females – the most prevalent area in females was the front chest above the umbilicus [
42]. In present study, a similar distribution of eschar was seen for both males and females except that eschars were more frequently detected in front chest above umbilicus in females than in males (Table
4). The distribution of eschar on body surface might be associated with dressing styles and personal hygiene, as the two factors affect how and where chiggers entered and stayed on the body surface [
5,
6,
9]. Ten eschars from 10 cases whose serum samples were IFA-positive and blood samples were PCR-positive were also tested positive by PCR. This result suggested that eschar could be used as an alternative indicator for the diagnosis of
O. tsutsugamushi infection [
27]. In addition,
O. tsutsugamushi DNAs were detected in all blood samples collected from 10 cases with eschar, indicating that the cases were bitten by scrub typhus infected chiggers.
Serodiagnosis methods of scrub typhus include Weil-Felix test, IFA, enzyme-linked immunosorbent assay (ELISA) and indirect immunoperoxidase (IIP). Weil-Felix test is neither sensitive nor specific, and replaced by IFA [
43,
44]. IFA is a well method for serodiagnosis of scrub typhus if it was conducted by skilled lab-persons. However, it always provides false positive results. Therefore, the real-positives might be lower than that obtained by IFA. ELISA and IIP are more accurate than the IFA [
45‐
47], but the two methods were not commercially available in China. Thus IFA was still widely used in China for serodiagnosis of scrub typhus until now. In present study, the confirmatory serodiagnosis of scrub typhus was made in case of a fourfold or greater rise in titers between paired acute and convalescent sera, or IgM or IgG titer in a single serum above 1:80, or 1:400 [
5,
24‐
26]. However, a few of cases whose IgM or IgG titer in a single serum under 1:80, or 1:400 might be neglected by IFA test according to this diagnosis criterion, especially when their convalescent sera were unavailable. This is a limitation of the present study.
The Kawasaki strain was first isolated from Japanese scrub typhus cases by using cyclophosphamide-treated mice in Miyazaki prefecture, and it was then found to be widely distributed in other infected areas of Japan, such as Kyushu, Kanagawa prefectures [
48,
49]. Kawasaki strain was defined as a new type of
O. tsutsugamushi distinguishable from prototype strains Gilliam, Karp, and Kato by antigenic analyses using polyclonal and monoclonal antibodies [
49]. However, some degree of cross-reactivity existed between the 54- to 56 kilodalton polypeptides of Kawasaki and Gilliam strains in immunoblotting analyses [
50]. Ohashi et al [
51] also found that the sequence homology of scrub typhus antigen 56-kilodalton (Sta56) surface protein gene between Kawasaki and Gilliam strain was the highest among the known
O. tsutsugamushi strains, and thus high cross-immunoreaction existed between the two antigens. Because Kawasaki strain was not available as antigen in China, we had to use Karp, Kato, and Gilliam strains as antigens for IFA. In present study, serotypes of IFA-positive sera were found to be the Gilliam type, while genotyping results of eschars, and isolates from some serologically confirmed cases suggested that the genotype to be similar to the Kawasaki strain. This inconsistency could be in part due to the existence of cross-immunoreaction between Kawasaki and Gilliam strain [
50,
51], such that antibodies to Kawasaki strain in the sera of cases in northern China could also be tested positive by IFA when we used Gilliam strain as antigen instead of Kawasaki strain.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
YL carried out the study design, epidemiological investigation, and drafted the manuscript. JS, YX, GL, LL, HX, NJ, YG, HY, SZ, PZ, JM, PF, SM participated in the epidemiological investigation, and sample and data collection in the field. DF, ZZ, WC and SL participated in the study design and helped to draft the manuscript. YL, SL, WC conceived the study and coordinated all the activities. All authors read and approved the final manuscript.